Over the past five years there has been steady development in the field of dynamic infrared scene generation (DIRSP), to the point where the first generation of dynamic infrared scene projectors can now be subjected to critical analysis. In this paper, the specific areas of optical projection validity, spectral emissivity and temperature resolution are discussed in relation to the viability of the DIRSP systems to which they apply.

The advent of missile seekers with dual-mode millimeter wave and infrared common-aperture sensors has led to a requirement to develop the simulation tools necessary to test these systems. Traditionally, one of the most important techniques for supporting systems development has been a full seeker hardware-in-the-loop simulation. The development of simulation facilities capable of supporting the new generation of advanced dual-mode guided systems has been limited due to some major technological challenges which are yet to be solved. This paper provides an overview of the development of such a simulation facility at the U.S. Army Missile Command for supporting hardware-in-the-loop simulations of dual- mode systems. The major technological challenges which limit common-aperture dual-mode simulator development are presented with the current approaches which are being taken to overcome these challenges. A description of the dual-mode simulator capable of providing simultaneous imagery to both sensors of a common-aperture seeker is provided.

The scene generation test capability (SGTC) has achieved an initial operating capability (IOC) at the Arnold Engineering Development Center (AEDC) using direct write scene generation (DWSG). This test tool will be able to present realistic mission scenarios directly to sensor focal plane arrays (FPAs) for developmental and operational test and evaluation (DT&E and OT&E), and will be integrated with the full-up sensor test capabilities at AEDC. The concept validation phase of this program is an operational system that is currently involved in sensor testing. The final phase provides scene projection at three infrared wavelengths and one visible wavelength. The facility is ready for FPA testing. This paper presents an overview of the current SGTC program, including a report of the hardware testing performed as part of the validation process.

The performance and analysis capability for earth observation systems (PACEOS) computer simulation software was developed by Martin Marietta Astro Space to assist in the evaluation of concepts and designs for space based remote sensing systems. This development took place over the last five years. PACEOS models the entire remote sensing process from the source of the radiation to the information required by the users. The simulation is divided into six primary sections which account for: (1) geometric corrections; (2) radiance sources and propagation; (3) sensor, detector, and platform properties; (4) data processing and transmittal; (5) data analysis; and (6) generation of user products, data, and images. PACEOS simulates the performance of earth and atmospheric remote sensing, surveillance, and early warning systems and is primarily used for concept definition and evaluation of hardware/software performance to evaluate the effects of sensor, platform, and processing parameters on the quality and accuracy of the information provided to the ultimate users. This paper discusses the procedures used to generate and propagate radiance from its source to the sensor aperture, including the effects of variations in surface albedo and temperature, illumination source and location, cloud layers and types, atmospheric models and parameters, cloud motion, target characteristics and location, spatial resolution, sensor bandpass, and geometric corrections. Microwave scene generation capabilities and other enhancements are being added to the current UV, visible, and IR capabilities under internal development efforts.

This paper describes recent technology at Arnold Engineering Development Center to provide real-time closed-loop image synthesis. Laser-based direct-write scene generation methods are used to simulate dynamic sensor operation and complex infrared scenes. New photonic image- synthesis methods employ image-to-object Whittaker-Shannon sampling, anisoplanatic optical convolution by quasi-isoplanatic spatial decomposition, and high-speed digital electronics for acousto-optic modulation.

This paper examines the relative significance of and dependencies between different noise sources which affect a sensor viewing a scene projector. An analysis is presented which compares the effect of various signal-dependent and signal-independent sensor noises, as well as the effect of projector nonuniformity (NU) on the sensor output. A key result of this analysis is a quantitative means to assess the importance of projector NU on sensor performance for different scene levels and operating conditions. It provides an analytical means to address questions such as: How and how much does projector NU influence the sensor response? At what level does the projector NU become a limiting factor in testing sensor performance? What is the penalty incurred in a particular test if a specified projector NU is not achieved? What effort should be expended to reduce projector NU in relation to other errors for a particular application? Discussion, results, and conclusions for a specific application are presented in addition to the analyses.

This paper addresses the process of measuring the output of individual elements of a pixelized scene projector. The in-band scene projector is a key component of a sensor/seeker test facility such as the Kinetic Kill Vehicle Hardware-in-the-Loop Simulator (KHILS) at Eglin AFB, Florida. Analyses are presented which quantify errors associated with measuring the radiant intensity of individual pixels on a scene projector. The errors are broken down into sampling errors, truncation errors, and random measurement noise. The magnitude of each error source is determined as a function of parameters of the projector and sensor such as the element spacings, and blur. Guidelines for using this information to accurately and efficiently perform nonuniformity correction of a scene projector are presented.

Kinetic energy weapon (KEW) programs under the Ballistic Missile Defense Office (BMDO) need high fidelity, fast framing infrared (IR) imaging seekers. As imaging sensors have matured to support BMDO, the complexity of functions assigned to the KEW weapon systems has amplified the necessity for robust hardware-in-the-loop (HWIL) simulation facilities to reduce program risk. The IR projector, an integral component of a HWIL simulation, must reproduce the real world with enough fidelity that the unit under test's (UUT) software will respond to the projected scenario of images as though it were viewing the real world. The CMOS resistor array IR scene projector, a wideband flickerless technology, shows great promise in cryogenic vacuum chamber as well as room temperature testing. A 128 X 128 CMOS resistor array has undergone extensive characterization measurements at Eglin AFB to determine its potential for HWIL testing of BMDO IR seekers. This paper addresses the nonuniformity correction (NUC) and use of the array in a calibrated projection test. The methodology and process for the NUC is described. Sensitivities to such things as output averaging, and optical sampling are explained. With the NUC procedure established, a test was accomplished that provided calibrated scene radiance values to a UUT. Absolute radiance values were not projected. Rather, the array's low and high output capabilities were equated to the low and high radiance values of an input scene. A calibration curve was established that allowed the UUT's output to be equated to the input scene's radiance values. The input scene was projected to the UUT, and the scene's radiance values were reproduced after applying the calibration curve to the UUT's output response. To the authors' knowledge, this if the first accomplishment of such a test with a dynamic IR scene projector.

The Optical Science Center, University of Arizona, has designed and constructed an optical system for infrared target simulator for Carco Electronics Inc. The system is a f/3.36 three mirror telecentric system with 12 cm aperture. Due to the working environment, finite element analysis was performed to assure the system quality.

This paper describes two aspects of work carried out at British Aerospace on a family of suspended resistor infrared scene generators intended as sources for exercising infrared seeker systems in simulation environments. In the first aspect, a 256 X 256 system has matured and entered service with hardware-in-the-loop (HWIL) simulation facilities. This system, designated TPS4 (for thermal picture synthesizer) has performance suitable for air target tracking studies, and certain aspects of its characteristics in use are described. In the second aspect, research work has been carried out on the extension of the system performance to enable the representation of higher temperature targets, such as are required for countermeasures work. These improved devices are designated TPS5, and aspects of their rationale, design, and evaluation are described. Prototype arrays suitable for eventual systems of complexity 512 X 512 and beyond have been tested.

A semiconductor tunneling infrared laser (STIRL) multiple quantum well device fabricated in GaAs/AlGaAs/InGaAs material is proposed as the emitting element of a two dimensional array for a scene projector. Such sources have a pico-second response time and hence can be modulated at up to 100 GHz rate, which may provide very high dynamic range to produce a sensitive gray scale. The STIRL device utilizes the intra-band emission from the excited state of a quantum well to produce radiation in the 4 - 15 micrometers range. It exploits a novel technique to inject electrons directly into the upper state of a quantum well structure by selective tunneling and allows them to relax to the ground state by emitting an IR photon and subsequently removing electrons from the ground state of the quantum well by another selective tunneling process. The energy difference between the lowest sub-bands in a quantum well of GaAs/AlGaAs can be as high as 0.25 eV which is equivalent to 5 micrometers emission wavelength. The quantum well dimensions can be tailored to produce emission in 5 - 15 micrometers . By selecting InGaAs for the quantum well materials and InAlAs for the barrier materials the emission range can be extended beyond these limits. Individual devices can operate at 100 to 200 mV and at low current resulting in power consumption of less than 15 mW depending on the size of the device. Monolithic integration of these microsources at high densities makes the high definition IR scene projector feasible.

generating a high degree of uniformity across the FOV in infrared scene simulators, over a wide dynamic range, is necessary to avoid introducing unintentional structure into the projected image. One challenge for calibration is establishing measurement of the radiance outputs of each of the 256,000 individuals pixels to the required accuracy levels at several radiance levels, within a reasonable time, with available instrumentation. Issues affecting measurement accuracy include the aperture, focal length, blur circle, and IFOV characteristics of the non-uniformity calibration (NUC) sensor, geometric and diffraction blur characteristics of the collimator optics (which vary with field position), NUC sensor noise and stability (temporal and spatial), emitter pixel geometry and temperature profile, and the relationship between the spectral characteristics of the NUC sensor and the source. Analyses are presented which determine the limitations on calibration accuracy based on predicted and measured performance of the WISP projector and the NUC sensor components. Some NUC sensor accuracy data, needed to support the determination of the overall process parameters, was collected in special NUC sensor tests and is presented herein. A combination of NUC process parameters is developed which achieves optimum accuracy in performing the NUC calibration, and which is expected to achieve the necessary calibrated uniformity performance of 1% for WISP.

The use of nonlinear infrared frequency conversion of laser output for infrared scene projection is explored. Nonlinear frequency conversion can be used to shift a laser wavelength for scanned laser images or used for two-dimensional image conversion in which an image in one spectral region such as the visible is projected as an image at a different wavelength in the IR.

High frame rate infrared scene generation depends on high performance digital processors that are tightly coupled to infrared emitter arrays. Massively parallel image generation hardware can realize the type of high throughput, high frame rate processing that will characterize the next generation of scene generators. This work outlines projects in massively parallel, high throughput image generation hardware using thin film optoelectronic devices which are integrated directly onto low cost silicon integrated circuits. For basic scene generation, an array of thin film emitters are placed on top of digital single instruction stream, multiple data stream (SIMD) parallel processors to provide high performance focal plane generation in a monolithic system. For more complex scene generation, low cost stacked silicon integrated circuits, using through-silicon wafer optoelectronic channels for three dimensional interconnections, form an extremely dense, high throughput, three dimensional parallel processing system. Thin film InGaAsP devices, which operate at wavelengths to which silicon is transparent, are integrated on top of standard foundry silicon integrated circuits so that stacked processor chips can communicate vertically. High speed analog interface circuitry on the Si integrated circuits provides a high bandwidth link between the devices and the digital processing circuitry. This processing approach provides tremendous generality for high frame rate image generation applications in a compact system. Issues addressed include system interfacing, power management, manufacturing tolerances, testing and repair, and system cost and effectiveness.

As IR sensor systems become more sophisticated, more stringent requirements are also placed on in-band IR scene projectors for hardware-in-the-loop (HWIL) applications. In general, the projector must be as good as, if not better than, the seeker in its performance. High scene fidelity, wide dynamic range, and high frame rates are now firm requirements, rather than idealized design goals. A novel type of projector system using diode lasers as sources offers several performance advantages over other currently available projector technologies. A projector of this type has been designed, built, and delivered to USAMICOM RDEC's HWIL branch under a Phase II Small Business Innovative Research (SBIR) award. The projector demonstrates high dynamic range, 128 X 128 scene size, and 8 KHz frame rates. The optical design for this system presents challenges not usually seen in projector systems, as it combines characteristics both of scanning and imaging systems. The optical system can best be described as a type of `reverse FLIR,' and uses anamorphic elements, an unusual polygon scan mirror, and a 2:1 telescopic relay.

A mosaic array of resistively heated microbridges offers flexibility for infra red scene simulations. The array may operate without flicker and display high-intensity dynamic scenes over a wide bandwidth. Honeywell completed fabrication of a 512 X 512 resistor array with 3.5 mils pitch for AEDC's 7V and 10V test chambers. The emitter has a broad bandwidth covering from 2 micrometers to 26 micrometers . The array operates at 20 K to simulate low radiation backgrounds in space. Up to 16,000 pixels may be turned on to simulate targets and target clusters. Each emitter element may heat up to 550 K with 1 kelvin resolution. The maximum power dissipation per pixel is 830 (mu) W for a pixel heated up to 550 K. The maximum power required is 13.2 watts for 16,000 pixels. This low power capability is derived from Honeywell's silicon nitride microbridge structure. Each emitter has approximately 85% fill factor and an average emissivity of 70% over the 2 - 26 micrometers bandwidth. Defect count in the array is less than 1% with one column out. The array may be addressed at 30 frames per second.

Several next-generation air-defense missiles will use dual-mode IR and rf guidance systems to address increasingly sophisticated threats. Performing hardware-in-the-loop tests of these systems to characterize IR and rf seeker performance, as well as developing and validating guidance algorithms, requires synchronously controlled IR and rf test targets, backgrounds, and countermeasures. The standard `electrically connected' approach uses separate IR and rf generation devices located in different laboratories and synchronized by a common control computer. This method has the disadvantage that the IR seeker must be physically removed from the guidance section and electrically reconnected via a long interface. The complex `collocated' approach combines IR and rf environment generators in the same enclosure, thereby eliminating the need for disassembly of the guidance section. A collocated IR/rf test facility is near completion at The Johns Hopkins University Applied Physics Laboratory. This article describes the new facility with emphasis on the IR environment simulator, which simulates targets, counter-measures, and background clutter.

A numerical model based on Monte Carlo simulation is proposed to study atmospheric turbulence effect on the received image by a remote sensing satellite. In the model, size and distribution of turbulent eddies, size of the subject on the ground and pixels on the receiver plane and wavelength of the optical beam are considered. Displacement of the pixels on the receiver plane is obtained and graphically shown.

The digital imaging and remote sensing synthetic image generation (DIRSIG) model emphasizes quantitative prediction of the radiance reaching sensors with bandpass values between 0.28 and 20.0 micrometers . The model embodies a rigorous end-to-end spectral modeling of radiation propagation, absorption and scattering, target temperatures based on meteorological history, extensive directional target-background interactions, and detector responsivities. This paper describes texture quantification, the spectral-spatial correlation of textures, texture collection and generation methods. Finally, we describe how DIRSIG generates texture on a pixel by pixel basis and maintains the spectral correlation of targets between bands.

The general scattered light (GSL) model is a computer program for predicting the spectral radiant intensity and in-band spatial radiance image for objects embedded in a scene. It was developed by Aerodyne Research under a Small Business Innovation Research project. Its significant capabilities include general bidirectional reflectance distribution function (BRDF), angularly structured illumination, general first surface reflections (via BRDF), infinite-order diffuse higher order reflections, estimated radiance fidelity, and optimized algorithms for short run-times. The general approach to radiance modeling taken in GSL makes it well suited for accurate object and background radiance modeling. GSL even includes the SEABEAM sea surface radiance model from Nichols Research for maritime applications.

Several mathematical models for the bidirectional reflectivity distribution function (BRDF) are compared to plane-of-incidence (PLIN), BRDF measurements of opaque surfaces. A laser at 0.6328 micrometers wavelength was the source for the BRDF measurement instrument; the theoretical measurement accuracy was to within 2%. Data are given for a glossy white paint, a glossy black paint, and a mill finished aluminum. The goal was to find a BRDF model that can represent the measured data to within 10% for any incident/reflected angle. Neither the Lambertian, Phong, Harvey, nor Cook-Torrance models had the desired accuracy for the three samples. A method which is based on interpolation and extrapolation of the empirical data is proposed which may achieve the desired goal both in and out of the PLIN.

Polarization characteristics of reflection and near-specular scattering by target surfaces at oblique incidence are evaluated by polarimetry in both infrared and visible wavelengths. Different materials with surfaces varying from smooth to very rough are investigated. Ellipsometry is used to measure the linear and circular polarizations for smooth to slightly rough surfaces. Polarized reflectance is measured for slightly rough to very rough surfaces. For specular reflection, dielectric surfaces show high linear polarization, and small circular polarization. Metal surfaces show high circular polarization, whereas small linear polarization and dielectric films show large linear and circular polarization. Near-specular scattering for a black anodized aluminum sample also shows polarization characteristics similar to those for specular reflection. Effective medium theory is used to model polarization for specular reflection, which shows agreement with the measured polarization for rough surfaces. Beckmann's scattering theory for random rough surfaces is used to model the near-specular scattering, which also shows agreement with the measured polarized reflectance. These data and the developed computation programs can be used to model polarization for targets of known geometric shapes.

Based upon known surface optical properties, we develop a model to calculate and analyze the linear and circular polarization signatures for targets of known geometric shapes. The linear and circular polarization radiation (emission and reflection) generated from the target surfaces are studied in two model surface structures: metallic and non-metallic substrates with/without dielectric coating. Infrared I, Q, U, and V images of a model cylindrical target with these surfaces are calculated. This paper shows that dielectric coating enhances the power of generating circular polarization radiation. In addition to the linear polarization, circular polarization imaging attributable to target surface reflection is also shown to be feasible for practical application.

The smart weapons operability enhancement (SWOE) program has developed a synthetic scene generation process that incorporates formal experimental design, random sampling procedures, data collection methods, physics models, and numerically repeatable validation procedures. The SWOE synthetic scene generation procedure uses an assemblage of measurements, static and dynamic information databases, thermal and radiance models, and rendering techniques to simulate a wide range of environmental conditions. The models provide a spatial and spectral agility that permits the simulation of a wide range of sensor systems for varied environmental conditions. Comprehensive validation efforts have been conducted for two locations: Grayling, Michigan and Yuma, Arizona, and for two spectral bands: shortwave (3 - 5 micrometers ) and longwave (8 - 12 micrometers ) IR. The intended use of the validated SWOE process is synthetic battlefield scene generation. The users of the SWOE process are the smart weapons system designers, developers, testers and evaluators, including developers of automatic target recognition algorithms and techniques.

Rapid advances in computer image generators (CIGs) and scene animating software facilitate the completion of important objectives in the distributed interactive simulation effort. One objective has been to produce an accurate simulation of meteorological visibility through the use of fog or blending functions. Silicon Graphics' (SGI) Performer software capabilities, for example, are shown here to be capable of correctly generating visibility and haze in real-time scenes if one applies certain physical definitions in conjunction with certain corrections to the SGI Performer software. This paper first presents a brief background on the physical basis of meteorological visibility. Then, it quantifies and tests the direct relationship between the scene generator's parameters for the `atmospheric fading coefficient' and the meteorological visibility. An important fix to the Performer's fog function is described and the results of our implementation of a meteorological visibility `calibration' are presented.

The Irma synthetic signature model was one of the first high resolution synthetic infrared (IR) target and background signature models to be developed for tactical air-to-surface weapon scenarios. Originally developed in 1980 by the Armament Directorate of the Air Force Wright Laboratory (WL/MN), the Irma model was used exclusively to generate IR scenes for smart weapons research and development. In 1988, a number of significant upgrades to Irma were initiated including the addition of a laser channel. This two channel version, Irma 3.0, was released to the user community in 1990. In 1992, an improved scene generator was incorporated into the Irma model which supported correlated frame-to-frame imagery. This and other improvements were released in Irma 2.2. Recently, Irma 3.2, a passive IR/millimeter wave (MMW) code, was completed. Currently, upgrades are underway to include an active MMW channel. Designated Irma 4.0, this code will serve as a cornerstone of sensor fusion research in the laboratory from 6.1 concept development to 6.3 technology demonstration programs for precision guided munitions. Several significant milestones have been reached in this development process and are demonstrated. The Irma 4.0 software design has been developed and interim results are available. Irma is being developed to facilitate multi-sensor smart weapons research and development. It is currently in distribution to over 80 agencies within the U.S. Air Force, U.S. Army, U.S. Navy, ARPA, NASA, Department of Transportation, academia, and industry.

The Ballistic Missile Defense Organization (BMDO) must simulate the detection, acquisition, discrimination, and tracking of anticipated targets and predict the effect of natural and man- made background phenomena on optical sensor systems designed to perform these tasks. NRL is developing such a capability using a computerized methodology to provide modeled data in the form of digital realizations of complex, dynamic scenes. The Synthetic Scene Generation Model (SSGM) is designed to integrate state-of-science knowledge, data bases, and computerized phenomenology models to simulate ballistic missile engagement scenarios and to support the design, development, and test of advanced electro-optical interceptor and surveillance systems. Multi-phenomenology scenes are produced from validated codes -- thereby serving as a traceable standard against which different BMDO concepts and designs can be tested. This paper describes the SSGM software architecture, the software modules and databases that are used to create scene elements, the synthesis of deterministic and/or stochastic structured scene elements into composite scenes, the software system to manage the various databases and digital image libraries, the ancillary software tool suite, and verification and validation by comparison with empirical data. The focus is on the functionality of the SSGM Release 6.0, and the planned development effort for subsequent SSGM releases.

Target range can be determined via triangulation with two IR sensors located along a known baseline. Achievable range accuracy depends on the baseline length, camera height, accuracy of relative orientation of cameras, and the accuracy of the angular position of the target. This paper investigates a method for determining the relative azimuth pointing angles of two IR cameras using a novel method in which sea background contrast features are used for registration. The method employs scene matching of sea contrast features with correction for optical distortion of the images and parallax. To evaluate this technique, IR scenes containing a target over the sea were acquired with two focal plane array IR sensors separated by about 30 m, and placed 30 m above sea level. Cross correlation functions were used for camera registration, and target range was calculated. Even with low sea-state conditions and low contrast irradiance, sufficient contrast features existed to generate meaningful cross-correlation functions. A requirement of the method is that the horizon must be accurately located in one of the two image pairs. The paper discusses experimental procedures, errors, data processing, and some results.

This paper describes recent research at Arnold Engineering Development Center to investigate aspects of using laser-based direct write scene generation photonics complementary to traditional blackbodies and thermal sources for focal-plane-array diagnostics. The equivalence of using lasers and thermal sources for photodetection and focal-plane-array evaluation depends not only on generating the same mean number of detectable photoevents, but also depends on generating the same photonic noise needed to accurately simulate photostochastic events representative of statistical ensembles and operational envelopes expected from normal focal-plane-array modes and sensor operation. Effective signal-to-noise ratios and photodetection equivalence also depend on the absence of signal artifacts or perturbations resulting from using laser-based photonics for optical diagnostics. The useful domain and relevance for using lasers complementary to thermal sources for photodetector and focal- plane-array diagnostics have been investigated to ascertain differences and similarities.

Optical techniques for space object identification have been a topic of research for several decades. Imaging techniques have been widely examined for this problem. However, it has been shown that high spatial resolution cannot always be obtained due to atmospheric turbulence. Spectral measurements are of interest to determine the material content when images cannot be obtained. In this paper spectra obtained from a Fourier transform spectrometer are used to identify materials using pattern recognition-based techniques. The signal-to-noise ratio of the Sagnac-type interferometer is derived, and the errors in material identification and abundance estimation that arise from using the measured spectra to estimate these quantities are studied. Fisher discriminants and multi-layer perceptrons were used to identify materials, and a constrained least squares technique was used to estimate abundances. Results for material identification and abundance estimation are presented as a function of signal-to-noise ratio.

The technique for calibrating color imagery which has been employed by the Tank-Automotive Research Development and Engineering Center (TARDEC) includes measurement of red, green, and blue color panels using a colorimeter during the approximate time that the calibration image is captured. This method has the advantage that the luminance and chromaticity coordinates of the color panels are recorded in real time. However, the disadvantage is the amount of time it takes to measure each individual panel. Outside of a laboratory, the environment cannot be controlled, so the light level and correlated color temperature from the source may shift during the calibration period. A new technique using a spectroradiometer has been developed whereby the spectral reflectance of the color panels are measured beforehand and only the light level and spectral content from the source is monitored during the calibration period. This drastically reduces the time required for calibration, thus rendering insignificant any temporal changes in the light level or correlated color temperature of the panels. The actual luminance and chromaticity of the color panels can be calculated subsequently.

An integrated approach to extract depth, efficiently and accurately, from a sequence of images is presented in this paper. The method combines the ability of the stereo processing to acquire highly accurate depth measurements and the efficiency of the spatial and temporal gradient analysis. As a result of this integration, depth measurements of high quality are obtained at speed approximately ten-times higher than that of the stereo processing. Without any a priori information of the locations of the points in the scene, the correspondence problem in the stereo processing is computationally expensive. In our approach, we use the spatial and temporal gradient analysis, which has been shown to provide depth with high efficiency but limited accuracy, to guide the matching process of stereo. Extensive experiments on real- scenes have shown the ability to acquire depth with mean error of less than 3 percent.

Rigorous functional integral equation of Dyson for average value and that of Bethe-Salpeter for correlation function of a wave scattered from a random rough Gaussian absolutely reflecting surface are derived on the basis of the Green formulas. Mass and intensity operators are not represented in an ordinary way as series or diagrams but as functional operators. The case of the infinitely small correlation radius is considered. In this case surface roughnesses with arbitrary heights have very steep slopes, and being reflected the waves can't but suffer the multiple scattering on roughnesses. Rigorous expression for an average reflected field is found, the mean surface being plane. It is shown that asymptotically the incident wave energy completely transforms into the coherent component of the field. This result is in accordance with the localization effect of the wave field in strong random media.

Sea radiance in the mid and far infrared shows a considerable degree of polarization which affects observed target-to-background contrast. An improvement in contrast is achieved with horizontal (s-plane) polarization filtering to suppress sea surface emission. Visibility and range affect the contrast in both polarizations. Scenes recorded during the MAPTIP measurement series off the coast of the Netherlands with the oceanographic ship HrMs Tydeman show decrease in contrast with range and better contrast for horizontal polarization against sea background. A simple mathematical model is presented relating contrast to extinction and path radiance which increase with increasing path length or worsening visibility.

Over the past several years, the author has reported to SPIE on several new instruments designed to provide important atmospheric measurements for electro-optical field testing. These include the millimeter wave transmissometer, the small portable infrared and visible transmissometer (SPOT), and the HOLODAR optical scintillometer. These instruments have reached either full operational status or have successfully completed the prototype phase. These instruments are described and examples of their data are presented. For the millimeter wave transmissometer, data showing performance in both rain and snow are presented with precipitation rate and drop size distribution and with comparisons to FASCODE calculations. For the SPOT, data from the visible and infrared regions are shown with comparisons with our older Barnes system and MODTRAN. The HOLODAR scintillometer data are compared to the NOAA/Lockheed scintillometer. Finally, the new trailers which house these instruments and their data processing systems along with the more standard meteorological data system are described. The resulting completely self contained mobile capability is available to support testing at any location.

Performance simulation and testing of weapons-based electro-optical sensors in the atmosphere often requires that naturally occurring and countermeasure aerosol particle size distributions be measured. Particle size distributions are necessary inputs for models used to predict the attenuation and scatter of radiation between the sensor and target. Current atmospheric measurement methods use optical individual particle counters or cascade impactors to sample a relatively small volume of air and to count or sort, by, equivalent scatter size or aerodynamic mass, aerosol particles. Particle measurement instrumentation is needed that is more cost- effective and more nearly reflects the particle size distribution along the entire propagation path of the sensor. The purpose of this paper is to describe a method for determining particle size distribution parameters from multispectral transmission measurements made along the line of sight of an electro-optical sensor under test. The limitations and errors associated with particle size distribution point sampling for atmospheric propagation paths are discussed and compared with those obtained from integrated path measurements. An algorithmic inversion of transmittance data to obtain particle size distributions, its limitations and capabilities is described. Example data inversion results are shown, and the application of the technique to the measurement to environmental air quality is discussed.

A diffuse target illuminated by the sun radiation is used for on-board calibration of a new stereo-spectral-imaging system ARGUS developed in the frame of the International project for MARS 94/96 missions. The target manufacturing technique is described. The experimental procedure used for getting the spectral and angular responses of target reflectivity is presented. The diffuse target was tested for the spectral properties at angle illumination -70 degree(s) (relative to the normal of the target) and angle of viewing +20 degree(s). A brief description of the setup for measurements of the spectral and angular responses of the target reflectivity is given. Results of the diffuse target calibration in the spectral range from 320 nm to 5200 nm are presented.

The FGAN-FfO has participated in the U.S. Smoke Week XV which took place in May 1993 at Eglin AFB, Florida. The primary objective was to study the effect of smoke clouds of various types on some selected electro-optical sensors. Two different thermal imagers, operating in the wavelength bands 2 - 5 micrometers and 8 - 11 micrometers , and two different laser rangefinders, at wavelengths 1.54 micrometers and 10.59 micrometers , were involved. The imagers were aimed against both tactical targets and point sources over a distance of about 600 m. The laser rangefinders were fired during the tests at 4 s intervals over tactical ranges against a laser reflection target. System performance of these devices in the presence of smokes was measured and analyzed. By combining the data of one of the imagers with that of one of the rangefinders, the effects of smoke clouds on a potential fire control system were studied. Comparisons between the statistical results show significant differences for the various smoke events in degrading the performance of the systems involved.

Observations of the open ocean have been collected at long range and low grazing angle with an airborne infrared system. These images are geographically registered using the position of a fiducial object in the scene as the reference. The three-dimensional frequency-wavenumber spectra computed from time histories of these images show a strong two-dimensional dispersion surface that is characteristic of wind waves and swell. The wave directions obtained from the spectra compare well with in situ measurements. Moreover, the wave speeds deduced from the spectra are consistent with the mean water depth in the area imaged. Ocean waves have been measured previously with other sensors, but these observations are believed to be the first measurements of ocean waves with an airborne infrared system.

An experimental design has been developed to facilitate collection of data for developing and testing computational models for assessment of the perceptual similarity or difference of texture patterns. This experimental design is based on Thurstone's Law of Comparative Judgement. To facilitate consistency in presentation of stimuli, collection of data, and computation of psychological scale values, an X-windows testing environment has been developed called the X-based Perceptual Experiment Testbed (XPET). A pilot study was conducted utilizing this experimental design. The study utilized images in which targets and their associated background had uncorrelated Gaussian noise texture patterns. Thus, only first- order image statistics were of significance. Psychological scale values for `target distinctness' obtained using this experimental design were compared to several first-order image metrics. Correlation coefficients as high as 0.9881 were found between the psychological scale values and first-order image metrics. It has been concluded that this experimental design should be adequate for data collection to support development of new second-order image metrics.

The detectability of a target in the infrared spectral region is determined by differences between the radiative signatures of the target and the local background. This implies that both, the difference in surface temperature and emissivity (Delta) T resp. (Delta) (epsilon) and the distribution of these differences over the target area and the background, are of major importance. Therefore camouflage measures have to address both issues in order to achieve maximum signature adaptation to the background. To determine the ability of a camouflage material to follow temperature changes in the background, temperatures measurements of camouflage systems and background elements have to be performed under a variety of meteorological conditions. Measurements of representative weather- and background conditions are needed to determine those situations where the camouflage material effectively reduces the target signature. The degree of temperature reduction depends on the required level of protection, that is for detection, recognition, and identification. Statistical analyses are given for various camouflage materials in relation to a number of background elements. Camouflage effectiveness is expressed in the percentage of time for which the apparent temperature contrast between the camouflage material and a background element is 1 degree(s)C, 2 degree(s)C, or 5 degree(s)C. Analyses are performed for five consecutive weeks of measurements in summer and winter, using data which were taken during a measurement campaign at Gilze-Rijen Air Force Base in 1990.

There is no question that Gauss developed the concept of least-square estimation which was stimulated by his astronomical studies. This concept was discribed in Gauss's book, Tlieoria Motus This contribution and insight provided by Gauss has inspired many researchers in estimation theory over the past 200 years. These developments include the Weiner Filter, Kalman Filter. Stochastic Estimation, Bayesian Estimation. Maximum Likehood Estimation, Auto-Regression and the Robust Filtering, just to name a few. However. during the recent decades, the need for detection and estimation of unknown signal in unknown noise background necessitated the development of correlation techniques for detection ( many correlation techniques were developed for identification). The problems in detection of unknown signals in unknown noise are common in ASW, ATR and in IRST images and ocean environment. Author's research in target detection in IR images and ocean environments let to his development of the "Correlation Filter". Correlation Filter became a part of his doctoral dissertation on a Generalized Filter where he has shown that all filters, Weiner, Kalman and Correlation Filters, are related through a "Constrained Gain Matrix" and that the Correlation Filter is a special case of the Weiner Filter, reference 2. This paper presents the derivation of the Correlation Filter for detection and estimation of unknown signals in unknown noise backgrounds and some applications. Reference I included two algorithms of his classified DoD applications. Since the paper has been selected for poster presentation, many photographs of the results of applications for this paper will be presented at the poster session.

The objective of a Battlefield Identification Friend/Foe-System (BIFF) is to make targets identifiable to the friend, the objective of camouflage is to make targets less conspicuous to the enemy. In the IR-spectral range these conflicting requirements can be met simultaneously to some extent by using material with variable radiance levels. The presentation addresses: (1) the different requirements posed by camouflage and BIFF, (2) the method of controllable energy radiation, (3) the approach to the solution of the camouflage and BIFF conflict, and (4) the limitation of the approach.

The theoretical and experimental research results of radar targets statistical characteristics at X- ad Ka-bands are presented. Data on radar cross-sections (RCS) for different targets including air, land, and marine targets are presented and the comparison of experimental results with theoretical models is done. The probability distributions of the signal instantaneous values, amplitudes, and RCS for different types of targets are obtained and the comparison of experimental distributions with standard Swerling's models is performed. The experimental results of power spectra investigations for different targets are discussed. For small marine targets the correlation between target radar spectra and sea surface spectra is determined.

Recently, the Naval Research Laboratory conducted 95 GHz passive and active measurements of a variety of targets on land and water backgrounds. The 95 GHz receiver system used stepped vertical raster scans to acquire images at both vertical and horizontal polarizations in passive mode, and at VV (vertical transmit/vertical receive), VH, HV, and HH polarizations in the active bistatic mode. Since the system operated in either passive or bistatic modes, the vertical and horizontal passive images are time-coincident, and with a single polarization source, the VV and VH bistatic images are time-coincident, as are the HV and HH images. Thus, comparisons among the various type of millimeter wave data are obtained by revisiting a given scene, and comparing areas within an image. The targets used were trihedral corner reflectors ranging in size from 20 to 60 cm, metal spheres, flat plates and sheets of microwave absorber. With 0.5 meter resolution at a range of 100 meters, and a 0.25 degree kelvin noise temperature for 0.1 seconds averaging, the system imaged these small targets floating on the water, and viewed scenes behind tree foliage. Initial analysis of the data using one-dimensional Fourier theory has enhanced the detection of the small targets. Further, additional spatial information obtained using polarization contrasts suggests the presence of sub-resolution elements within the images that may be enhanced using beam-shaping techniques.

High spectral resolution observations of several common camouflages have been made as part of a utility study for a Fourier transform visible hyperspectral imager (FTVHSI). Two types of data were collected. First were non-imaging spectral reflectance measurements made with a spectral resolution better than 0.3 nm over the 350 nm to 1150 nm band. Second were hyperspectrally resolved two dimensional hypercubes of the samples using a FTVHSI. These second data have a spectral resolution of 270 cm^-1 over a band of 370 nm to 1030 nm and a spatial resolution of about 2 cm. The data were taken against representative foliage backgrounds that ranged from grass, to tropical forest vegetation, to an arid mesa. The data show both macro and micro spectral differences between the camouflage and the backgrounds that are apparent in the hyperspectral renditions but missing in broad band observations.

A series of field measurements of targets and backgrounds was made by an infrared Fourier transform spectrometer as part of the Joint Multispectral Sensor Program (JMSP), a joint program involving the U.S. Navy, U.S. Air Force, U.S. Army, and ARPA. These measurements were designed to observe targets in various types of background clutter, investigate the utility of novel algorithms for the detection of resolved and sub-pixel targets, and to aid in the selection of spectral bands for a future airborne multispectral sensor. This paper gives an overview of objectives and goals of the JMSP data collections, the targets viewed, and examples of the observed data.

In this paper, the incorporation of atmospheric aerosol and turbulence blur and motion blur into visible, near infrared, and thermal infrared target acquisition modeling is considered. Here, we show how the target acquisition probabilities and, conversely, the ranges at which objects can be detected are changed by the inclusion of these real-life environmental effects whose blur is often significantly greater than that of imaging system hardware. It is assumed that images are contrast-limited rather than noise-limited, as is indeed the case with most visible, near infrared (IR), and thermal IR sensors. For short focal lengths with low angular magnification, such environmental blur effects on target acquisition are negligible. However, for longer focal lengths with large angular magnification, resolution is limited by them and this has a strong adverse effect on target acquisition probabilities, times, and ranges. The considerable improvement possible with image correction for such environmental blur automatically in a fraction of a second is significant for contrast-limited imaging, and is discussed here too. Knowledge of such environmental MTF is essential to good system design and is also very useful in image restoration for any type of target or object.

The search procedure and target acquisition in cluttered scenes are highly dependent on the contrast between target and background, and the scene content. The latter varies widely from uniform displays of sky or sea to highly complex displays of mixed landscapes or urban areas. Generally, contrast should be defined within the limit of the geometrical resolution of the sensor system. Both target and background will be structured in respect of optical radiance or thermal emittance. Lillesaeter has presented a contrast definition allowing for structured as well as plain targets and backgrounds. It consists of two parts, area contrast and edge contrast, both of which are logarithmic functions of target/background radiance ratios. This contrast definition has been implemented in the Night Vision Laboratory (NVL) model to simulate acquisition of thermally structured targets in complex backgrounds. As far as the effect of clutter makes the task of discriminating the target from background more difficult, videosimulations have been used to correlate the NVL model to different levels of scene complexity. This paper presents the NDRE modification of the NVL model and gives some examples related to the efficiency of thermal camouflage.

Over the past several years there has been a large quantity of infrared target and background signature data collected using imaging radiometers. As with any data, one of the primary questions that must be answered is, `How has the measurement system affected the data?' During the smart weapons operability enhancement (SWOE) signature measurement program this question was addressed using two `identical' AGEMA 880 dual band radiometer systems. These two systems were used to determine how much of the variation that is observed in the signature data can be accounted for by measurement system differences. Measurements were made at the Yuma Proving Ground over a 47 day period from 15 March to 30 April 1993. The radiometers were mounted side-by-side and the fields of view registered as closely as possible. Most of the background signature measurements were preceded by simultaneous blackbody measurements with all four radiometers. These blackbody measurements were used to compare the performance of the two systems under the various weather conditions experienced during the measurement period. The blackbody measurements have shown that over an ambient temperature range of almost 40 degree(s)C the variation between radiometer pairs was less than 1 degree(s)C for the long-wave (8 - 12 micrometers ) and less than 2.5 degree(s)C for the short-wave (3 - 5 micrometers ).

A series of infrared hyperspectral field measurements was made at Wright Patterson Air Force Base and the U.S. Army White Sands Missile Range by the Joint Multispectral Program (JMSP) between November 1993 and June 1994. In these experiments, a highly sensitive Fourier transform spectrometer (FTS) was used to collect data from test panels, military and civilian vehicles, and various types of natural backgrounds. Post-collection data analyses are being conducted by the JMSP to assess the potential of thermal multispectral processing for detecting and classifying low-contrast ground targets in natural clutter environments. One target material of special interest is CARC paint, which is currently applied to U.S. Army vehicles in various colors to create woodland and desert camouflage patterns. CARC-painted test panels were observed in a wide variety of backgrounds and weather conditions during all of the JMSP experiments. It is shown here that certain fine-scale spectral features of this paint can support reliable two-color discrimination of CARC-coated test panels in different natural backgrounds, even under low contrast and high clutter conditions. The paper also examines environmental variations in two key parameters that determine spectral detectability; specifically, the observed target-background spectral contrast signature (which provides the required coloring), and the background spectral correlation (which provides for multiband clutter suppression).

This paper describes an electromagnetic computer prediction code for generating radar cross section (RCS), time domain signatures, and synthetic aperture radar (SAR) images of realistic 3-D vehicles. The vehicle, typically an airplane or a ground vehicle, is represented by a computer-aided design (CAD) file with triangular facets, curved surfaces, or solid geometries. The computer code, XPATCH, based on the shooting and bouncing ray technique, is used to calculate the polarimetric radar return from the vehicles represented by these different CAD files. XPATCH computes the first-bounce physical optics plus the physical theory of diffraction contributions and the multi-bounce ray contributions for complex vehicles with materials. It has been found that the multi-bounce contributions are crucial for many aspect angles of all classes of vehicles. Without the multi-bounce calculations, the radar return is typically 10 to 15 dB too low. Examples of predicted range profiles, SAR imagery, and radar cross sections (RCS) for several different geometries are compared with measured data to demonstrate the quality of the predictions. The comparisons are from the UHF through the Ka frequency ranges. Recent enhancements to XPATCH for MMW applications and target Doppler predictions are also presented.

Accurate models of millimeter-wave (MMW) radar ground clutter allow improved evaluation of algorithms that detect and identify military vehicles for smart weapons systems operating in a clutter environment. To this end, an empirical model was developed that uses radar clutter data measured at 95 GHz for desert and European-like environments measured over approximately 6 to 7 week periods. Using low-angle, frequency-averaged MMW radar cross sections per unit area ((sigma) ⁰) data that were collected in the spring season at Yuma, Ariz. and Grayling, Mich., I evaluated a multivariate linear regressive model of (sigma) ⁰ based upon least-square estimation using measured environmental parameters as independent variables. The results for the data collected at Yuma indicate that (sigma) ⁰ values were dependent upon the parameter of chronological time, and moderately dependent upon the parameters of soil moisture content and relative humidity. The results for the data collected at Grayling indicated that (sigma) ⁰ values were highly dependent on the environmental conditions. The model successfully described refrozen ground and snow, and drying ground conditions using the available environmental data, but melting, transitional, and falling snow conditions were not successfully described. For the entire Yuma data set and for a large portion of the Grayling data set, environmental parameters were identified and incorporated into a linear model that described the variations in (sigma) ⁰.

A high-speed dynamic IR scene projector based upon diode lasers has been designed, fabricated, and delivered to the U.S. Army Missile Command's (USAMICOM's) Research, Development, and Engineering Center (RDEC). The projector was developed under a Phase II Small Business Innovative Research award. The projector is based upon a linear array of Pb- salt diode lasers coupled with a high-speed optical scanning system, drive electronics and synchronization electronics. The projector is capable of generating high dynamic range, 128 X 128 scenes at 8 KHz frame rates. The system's modularity provides upgradability to meet specific performance requirements such as increased spatial resolution, different emission wavelengths, or dual-band scene projection. The projector's performance characteristics are presented in the paper, as well as sample images generated with the projector and captured by an InSb FPA sensor.

The multi-spectral sensor calibration system (MSSCS) is a highly automated calibration system designed around an optical table. It consists of several instruments and sources mounted on two translating tables and placed at the focus of an off-axis parabolic collimating mirror. These instruments and sources are linked to a control computer via the IEEE-488 buss. Sophisticated system control software consists of interlinked programs that provide: (1) instrument control and data acquisition, (2) automated calibration and data analysis, and (3) automated report generation. The MSSCS is used to calibrate several of NASA's airborne multi-spectral sensor systems currently in use to provide remote sensing data. These systems include the CAMS (calibrated airborne multi-spectral sensor), the ATLAS (airborne terrestrial applications sensor) and TIMS (thermal infrared multi-spectral scanner). We begin by briefly describing the overall system design with an emphasis on the objectives of the design and the capabilities provided by the complete system. This is followed by a description of how the system is aligned with the unit under test. Examples of data produced by the system are presented, and how this data was certified correct. We conclude with how the system's data products were validated.

A fast approximation method for simulating the dynamic effects of atmospheric turbulence on infrared digital imagery is presented. The method is based on the distant phase screen approximation for certain turbulent conditions and is an application of the theoretical work published by M. I. Charnotskii, et al. in 1990. The computational efficiency of the method allows realistic incorporation of the atmospheric turbulence effects in high resolution interactive simulation environments for military and commercial applications. Simulations of turbulence effects on digital infrared images are presented together with actual images obtained from a commercial 8- to 12-micrometers infrared system imaging through thermally induced atmospheric turbulence.